Statement of the Problem: Rates of youth violence perpetration and victimization vary dramatically across neighborhoods, contributing to health disparities within the U.S. across race and social class lines. Purpose of the Proposed Project: This project seeks to evaluate the effects on youth violence perpetration and victimization of a randomized housing policy intervention in the city of Chicago designed to de- concentrate areas of poverty and violence. The use of random assignment to the intervention wait list together with administrative data enables us to carry out an incredibly cost-effective evaluation that is as rigorous as any randomized experiment - but at just a small fraction of the cost. Methods and Study Population: In 1997, Chicago re-opened the housing voucher program's wait list for the first time in 12 years. A total of 82,607 eligible families applied, far in excess of the number of available vouchers, and so applicants were assigned by random lottery to a wait list. We focus here on the 10 percent of families living in public housing at baseline. Below we provide preliminary results showing that for these families, vouchers generate sizable changes in community environments. On the other hand wait-list position is unrelated to baseline socio-demographic characteristics - that is, the random voucher lottery was indeed random. We measure youth involvement in violence perpetration by matching to administrative arrest records from the Illinois State Police (ISP), and measure youth violence victimization using mortality data from the National Death Index (NDI). The causal effects of housing vouchers on youth violence outcomes are identified by essentially comparing outcomes for youth whose families are randomly assigned to "good" versus "bad" wait-list positions. Evidence on how effects vary across youth by baseline characteristics can provide information about how community-level risk and protective factors interact with youth-level risk and protective factors such as age, gender, and family structure, as well as developmental factors such as prior offending trajectory. Information about mechanisms will come from analyzing voucher impacts on individual- and community-level candidate mediating factors. Finally, outcomes will be compared for youth whose families make different types of moves with their housing vouchers to examine whether neighborhood effects on youth violence are linear or non-linear with respect to specific neighborhood attributes. Implications for Prevention: The findings from our project should be relevant to a vast set of housing policy decisions that have the potential to shape the concentration and overall volume of youth violence in America. Our study may also directly aid in the design of new community-oriented health interventions targeted at the specific factors within the social environment that influence youth violence. This project seeks to evaluate the effects on youth violence perpetration and victimization of a randomized housing policy intervention in the city of Chicago designed to de-concentrate areas of poverty and violence. The use of random assignment to the intervention wait list together with administrative data enables us to carry out an incredibly cost-effective evaluation of an intervention that helps youth move to less disadvantaged, dangerous neighborhoods that is as rigorous as any randomized experiment - but at just a small fraction of the cost. Findings from this study should be relevant to a vast set of housing policy decisions that have the potential to shape the concentration and overall volume of youth violence in America, and may also directly aid in the design of new community-oriented health interventions targeted at the specific factors within the social environment that influence youth violence.